See the WebSphere eXtreme Scale Wiki for links to eXtreme Scale Version 7.0 documentation.
If you log in with your developerWorks ID, you can leave comments and feedback for the development team.

IBM WebSphere eXtreme Scale Version 6.1 Fix Pack 3 (6.1.0.3), including the ObjectGrid component, provides fixes, as well as several new features and function enhancements. Version 6.1.0.3 supersedes ObjectGrid 6.1.0.2 cumulative fix 2.

Download Information

Download WebSphere eXtreme Scale Version 6.1.0.3 from the IBM support site. For more information, see 6.1.0.3: WebSphere Extended Deployment V6.1 Fix Pack 3 for multi-platforms .

Complete the following steps to install WebSphere eXtreme Scale Version 6.1.0.3:

1. Install IBM WebSphere eXtreme Scale Version 6.1
2. Download and install 6.1.0.3: WebSphere Extended Deployment V6.1 Fix Pack 3 for multi-platforms

Features and enhancements

ObjectGrid includes the following new features:

Programming model enhancements

• EntityManager API
  The following EntityManager API enhancements are available:
  • Enum support
    Entities have first-class support for enum data types, which are available in Java SE 5 and later.
  • Entity merge support
    Detached entities can be merged into ObjectGrid using the EntityManager.merge(Object entity) method, allowing simplified, automatic entity inserts and updates.
  • Entity invalidation support
    Entities can be invalidated from ObjectGrid without affecting the loader using the EntityManager.invalidate(Object entity) method.
• Query API
  The following Query API enhancements are available:
  
  • Simple plan retrieval
    Query plan retrieval is simplified using the Query.getPlan() and ObjectQuery.getPlan() methods. Prior to 6.1.0.3, the query plan could be displayed only by enabling trace.
  • Enum support
    Enum data types in ObjectMap objects or entities can now be queried using parameters or literal strings.
• ObjectMap API
  The following ObjectMap API enhancements are available:
  • Map clear
    Maps can be efficiently cleared using the ObjectMap.clear() and JavaMap.clear() methods. This method avoids the need for costly map entry iteration. Read more...
• System API
  The following System API enhancements are available:
  • State manager
    The ObjectGrid StateManager interface allows administrators to mark the grid off-line. Client operations are quickly rejected when the ObjectGrid is in off-line state. Clients also have the ability to check the state of the grid. Read more...
  • Memory aware evictors
    All of the built-in ObjectGrid evictors can be configured to evict elements from maps if memory utilization thresholds are reached. This decreases the chances of running out of memory in many scenarios. Read more...
  • Session handles
    When using per-container placement, clients can retrieve a handle to the session where a request is routed. For subsequent requests, the handle can be reapplied to the session, providing affinity to the same session.
  • Proximity routing
    Clients can be configured to communicate with ObjectGrid servers that are closest to them. If the server is in the same process, machine or zone, the client can automatically route replica read requests or requests to per-container placement maps to servers that have less network latency.
  • Heartbeat tuning
    There are three heartbeat settings that can be used to tune how aggressive ObjectGrid will monitor the availability of the containers. Read more...

Database integration

• JPA cache plugin
  ObjectGrid includes an OpenJPA and Hibernate cache plug-in that allows ObjectGrid to be used as the main side cache for OpenJPA and Hibernate. Read more...

• JPA loader
  The JPA loader plug-in allows simplified integration with database back-ends. Both OpenJPA and Hibernate can be used to automatically convert relational data to and from Java objects or ObjectGrid entities. Data can be automatically loaded into ObjectGrid and can keep the back-end database in sync without writing complicated JDBC code. Read more...

• Write-behind caching
  Updates to a loader can now be automatically buffered and asynchronously written to the back-end. This can decrease the load to the back-end significantly by batching and merging the updates together and improve application performance by removing the back-end write from the application transaction. The buffer is replicated within ObjectGrid for high availability. Read more...

• Automatic stale data invalidation
  ObjectGrid includes a plug-in that automatically keeps ObjectGrid updated with database changes. This can be done automatically by using the System Change Number (SCN) support included with Oracle 10g or IBM DB2 9.5 or by using a user-defined timestamp field in the database. Read more...

Bug fixes

• Client replica stops replicating after server failover — A client replica does not receive new data after a server fails.
• Duplicate entities may not be detected when using CascadeType.PERSIST — Duplicate entities may not be detected correctly when using CascadeType.PERSIST.
• Entity queries returning entities are very slow — Queries that return entities cause multiple remote calls to the server.
• EntityManager does not correctly detach lazy entities — Unfetched lazy entities can be incorrectly fetched after the parent entity is detached.
• Exceptions occur in Catalog Server when stopping many servers at once — When stopping large numbers of ObjectGrid servers at the same time, the catalog server can get exceptions.
• Property access entity updates unnoticed when using Java instrumentation agent — The dirty bit in the entity proxy is not set when calling setter method within the same entity instance.
• Several problems with EntityManager temporal attributes — Dates and times are not correctly normalized and nanoseconds is lost for Timestamp types.

Wiki Disclaimer and License

© Copyright IBM Corporation 2007,2009. All Rights Reserved.